LearnProduct data

The product information chain: from PIM to page to feed

By Published 3 min read

Explore with AI

The short answer

The product information chain is the path a product fact travels from the source system through a PIM, a catalog, a web page and an outbound feed to the reader. Each hop can transform the value: a unit converted, a field truncated, a revision dropped. Accuracy is a property of the whole chain.

In this article · 6
  1. The path of one fact
  2. What each hop can change
  3. Find the first wrong hop
  4. Carry the source through every hop
  5. Ownership by hop
  6. Questions

The path of one fact

A specification begins in an engineering system or a manufacturer data file. From there it is typically loaded into a PIM or an ERP, mapped into a catalog structure, published to a website page, exported into feeds for distributors and marketplaces, and finally read by a buyer, a search engine or an assistant. At every hop someone or something decides how the value is represented.

The chain is long, owned by different teams, and rarely drawn on one page. That is why an error found at the end is hard to attribute to the hop that caused it.

What each hop can change

  • Load: a spreadsheet or file import that reads a unit as text, drops leading zeros or converts a fraction.
  • Mapping: a source field with a basis (nominal or actual) mapped to a destination field that has none.
  • Template: a web page that concatenates attributes into a sentence and truncates at a character limit.
  • Feed: an export that rounds, renames or reformats values for a marketplace's schema.
  • Publish lag: a correction made at the source that reaches the page in days and the feeds in weeks.
  • Cache: a copy retained long after the source changed.

Find the first wrong hop

Take a wrong value found at the end and walk the chain backward. Compare the value at each hop with the hop before it. The first place where the two differ is where the error entered. Everything downstream is propagation. The comparison needs the intermediate copies to be retrievable, which is a design requirement of the chain, not a debugging convenience.

Carry the source through every hop

A value that moves through the chain should carry, with it, its source, its revision, who confirmed it and when. Most transformations discard these because the destination field has nowhere to put them. The consequence is that the reader cannot tell a value confirmed this week from one copied three years ago.

The remedy is to treat the source as part of the value: either a field alongside it in each hop, or a record outside the chain that each hop points to. The second is more practical across organizations, because it does not require every system to change its schema.

Ownership by hop

Each hop has an owner. The data steward owns the load. The merchandising team owns the template. The e-commerce platform owns the feed. Without an agreed record of what the value is supposed to be, each owner defends their hop as correct and the error belongs to nobody. The checked record is the shared reference all of them compare against. It is the subject of what is Product Truth and Product Truth versus a PIM.

Questions

Is the chain the same as a PIM?

No. A PIM is one system inside the chain. The chain also includes the source data, the catalog, the web pages, the feeds and the readers.

How do I find which hop introduced an error?

Compare the value at each hop with the one before it, starting from the wrong value at the end. The first difference is the entry point.

Why does a correction take so long to appear everywhere?

Each hop publishes on its own schedule and caches its own copy, so a source correction reaches pages and feeds at different times.

What should travel with a value?

Its source, revision, the person who confirmed it and the date, either as fields or as a pointer to a record outside the chain.

Explore with AI

Where does your own information disagree?

Get Upstream

One email a week: what we asked the AI, what it got wrong, and how we checked.

Our research reaches 17,000+ followers across our channels. Upstream is the weekly version.

Weekly. Unsubscribe anytime. Privacy